From: | Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com> |
---|---|
To: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Consolidate 'unique array values' logic into a reusable function? |
Date: | 2016-08-07 04:45:39 |
Message-ID: | CAEepm=2vmFTNpAmwbGGD2WaryM6T3hSDVKQPfUwjdD_5XY6vAA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi,
Looking at commits f10eab73d and c50d192c, I wondered why we don't
have a reusable in-place unique function. It may be trivial, but we
seem to have a lot of copies and variations in the tree.
Here's a sketch patch that creates a function array_unique which takes
the same arguments as qsort or qsort_arg and returns the new length.
The patch replaces all the specialised unique functions and open coded
versions that I could find with simple greps, but there are probably
more.
My compiler seems to inline the comparator function and memcpy well,
so I can't measure any speed difference between array_unique(array,
size, sizeof(int), compare_int) and a hand-crafted loop using == for
comparison and = for assignment, for a billion items.
If no one objects I'll post a version of this to a commitfest, along
with some other trivial code duplication refactoring work I posted a
while back that consolidates popcount and ffs/fls implementations. I
don't like code duplication :-)
--
Thomas Munro
http://www.enterprisedb.com
Attachment | Content-Type | Size |
---|---|---|
array-unique.patch | application/octet-stream | 16.7 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Dean Rasheed | 2016-08-07 07:01:40 | Re: Bogus ANALYZE results for an otherwise-unique column with many nulls |
Previous Message | Tom Lane | 2016-08-07 02:13:16 | Draft release notes for next week's back-branch releases |